International experts at war against AI?

AI technology is advancing by leaps and bounds and revolutionizing many fields. However, it comes with numerous disadvantages. Deepfakes are among the dangers of this technology. In the face of the proliferation of these fake digital contents, experts are sounding the alarm and calling for stricter legislation.

An appeal for strict regulation to protect digital integrity

Current artificial intelligence legislation does not sufficiently regulate the production and dissemination of deepfake content. This is the complaint of a group of international experts in an open letter entitled “Disrupting the Deepfake Supply Chain” published on February 21st. These researchers invite legislators to adopt strict measures to put an end to the distribution of deepfake contents, notably images and videos of a sexual nature or aimed at political misinformation.

In the group of experts, we find thought leaders such as Andrew Yang, Stuart Russell, Joy Buolamwini, Sarah Gardner, Gary Marcus, Steven Pinker, Sneha Revanur, and many other influential personalities. In this letter, the experts propose concrete recommendations to limit the multiple negative consequences of these AI-generated contents.

They urge legislators to criminalize pornographic deepfakes and to provide penalties for the creators of this harmful content. The experts also demand that AI software developers and distributors take action to prevent the creation of these deepfakes.

Les deepfakes p0rn0 de Taylor Swift enflamment Internet depuis quelques jours… et les journalistes sont en train de blâmer les nouvelles IA ! 👀

Mais en réalité, les IA savent faire ça depuis longtemps :

L'une des 1e vidéos deepfake date de 1997 et peut se retrouver sur… pic.twitter.com/ao4e6YdLZu

— Lou (@byLouIA) February 17, 2024

A response to the proliferation of these fake recordings generated via AI

Deepfakes have literally exploded in 2023. The Home Security Heroes website estimates over 95,820 deepfake videos generated via AI and broadcast over the past year. This represents an increase of more than 550% compared to the numbers in 2019. Pornography makes up a significant proportion of this content, accounting for 98% of the deepfake videos generated in 2023 on its own.

In the United States, for example, several deepfake contents have flooded social media in recent weeks. False pornographic images of the famous singer Taylor Swift first stirred the X social network. The published content was viewed by over 45 million netizens and registered hundreds of thousands of likes. Another pornographic video of certain high school students from New Jersey generated with the aid of AI also went viral on social networks.

Finally, an audio deepfake impersonating the voice of President Joe Biden was made to prevent Democratic voters in the state of New Hampshire from turning out to the polls. These events underline the urgency of the situation and the importance of legislation that strongly condemns the production and dissemination of these fake contents created with the help of AI.